CS194 Project 6A

Image Warping and Mosaicing: Part A
by Vanessa Ng

1: Shoot and digitize images

2: Recover Homographies

I used matplotlib.pyplot.ginput to supply four correspondence points between the source and the 'centered' image.

The goal is to compute the homography matrix, H:

To do so, we set up the following equations in AH = b format:

I set up a linear system of several equations and use numpy's least square solver to solve the best estimate.

3: Rectified Images

I warped the images using skimage.transform.warp function. I supplied the source image and H inverse to the function.

4: Mosaic

Image mosaics were created by warping the source image to a 'center' image perspective.

Original images

Warped and combined

Original images

Warped and combined

Original images

Warped and combined

Original images

Warped and combined

What I learned

I learnt to better visualise and understand how panaromas are created!